COPYRIGHT (C) 1984-2021 MERRILL CONSULTANTS DALLAS TEXAS USA
MXG NEWSLETTER SIXTY-FIVE
***********************NEWSLETTER SIXTY-FIVE***************************
MXG NEWSLETTER NUMBER SIXTY-FIVE - DATED Jan 6, 2015.
Technical Newsletter for Users of MXG : Merrill's Expanded Guide to CPE
TABLE OF CONTENTS
I. MXG Software Version.
II. MXG Technical Notes
III. MVS, aka z/OS, Technical Notes
IV. DB2 Technical Notes.
V. IMS Technical Notes.
VI. SAS Technical Notes.
VI.A. WPS Technical Notes.
VII. CICS Technical Notes.
VIII. Windows NT Technical Notes.
IX. z/VM Technical Notes.
X. Email notes.
XI. Incompatibilities and Installation of MXG.
See member CHANGES and member INSTALL.
XII. Online Documentation of MXG Software.
See member DOCUMENT.
XIII. Changes Log
Alphabetical list of important changes
Highlights of Changes - See Member CHANGES.
COPYRIGHT (C) 1984,2015 MERRILL CONSULTANTS DALLAS TEXAS USA
I. The 2015 Annual Version MXG 32.32 was dated JAN 6, 2015.
You can always use this form
http://www.mxg.com/Software_Download_Request
to request the ftp download instructions for the current version.
ALWAYS USE THE MOST CURRENT VERSION OF MXG.
See CHANGES member of MXG Source, or http://www.mxg.com/changes.
II. MXG Technical Notes
1. Text:
III. MVS, a/k/a z/OS, Technical Notes.
19. APAR PI28227 reports incorrect measurements of DB2 DDF, and a memory
leak in DB2 SMF IEFU84 that occurs only when SMF record compression
is enabled (DB2 V10 or higher). In this case, the DB2 DDF address
space was the victim; it ABENDed and brought down the DB2 subsystem.
18. APAR OA45637 reports SYSPLEX-WIDE hand caused by SMF prolonged use
of the CMSSMF lock while recovering from a data lost condition. The
APAR changes data loss recovery processing to pre fetch the storage
containing the BQEs prior to obtaining the CMSSMF lock, to reduce
the time required for paging while holding the lock.
17. APAR OA46206 reports SMF 82 subtype 9 records are not written when a
a key is deleted from the CKDS; subtype 9s are written when a key is
added or changed.
16. APAR OA45117 reports SMF 42 subtype 18 might not be written after a
termination or restart of SMFSVSAM on the first system, when the SMF
write duty is passed to a second system.
15. APAR OA45348 reports CF CPU utilization increase after activating
SM Duplex for MQ Application str, due to excessive suppression for
latch contention during system-managed duplexing, plus details on
how CF duplexing actually works referencing RMF I reports.
14. APAR PI24782 reports IBM HTTP Server V8.5 on z/OS does not write
SMF type 103 subtype 14 unless SMFLogDebug ON is set, even though
mod_smf was correctly set.
13. APAR PI18028 reports a DB2 SINGLE INSERT took 442,830,000 GETPAGES
for a small classic segmented tablespace, when using DB2 TRUNCATE
statement operating on classic segmented or Universal Table space.
12. APAR OA46206 reports SMF 82 Subtype 9 records are not written when a
KEY is deleted from the CKDS; records are created only when keys are
added or changed.
11. APAR OA46156 reports PFA Fault Analysis PFA_SMF_ARRIVAL_RATE misses
some SMF records in some SMFPRMxx configurations.
When a customer specifies a SUBSYS(STC...) in their SMFPRMxx member,
PFA's dynamic exit that listens for SMFs arriving, does not listen
for the SMF records for STCs. PFA_SMF_ARRIVAL_RATE only has dynamic
exits for the records defined in SMFPRMxx for SYS. This can result
in SMF records to not be counted, which may prevent jobs from being
monitored in PFA's activity
10. APAR OA45118 reports SMF 42 Subtype 18 may not be recorded after
a termination or restart of SMFVSAM on a different system.
9. APAR OA46331 reports too many SMF Type 80 RACF records are written
for resources listed in multiple Grouping Class profiles, filling
SMF datasets.
APAR PI29659 reports excessive SMF 80 records with LOG=NOFAIL when
auditing the AUTH REQUEST (with DB2 Auth Exit DSNX@XAC enabled.
8. APAR OA39062 reports for z/HPF enabled devices, SMF 42 Subtype 6
records contain incorrect disconnect time for READS (S42DSRDD) and
incorrect read operations (S42DSRDT). OA31856, OA39169 related.
7. APARs OA44491 and OA45373 correct IP Print Way SMF 6 records, in
particular, SMF6NLR/NRLINES and SMF6NDS/NRDSETS were always zero
and some flags in SMF5DCI/DCI1/CONTRIN0-3 were wrong.
6. Measuring the cost of SMS Hardware Compression - some thoughts from
a Hardware engineer:
The CPU cost is recorded in the SMF 30 records for the task, so you
can use SMF 30 records for comparisons of using and not-using, but
there are no separate metrics for the cost of compression itself.
The LZW algorithm used is asymmetric: compression uses more CPU
than decompression (a/k/a compaction/decompaction to IBM).
Comparisons were made with some simple IEBDG patterns, then with
files full of random floating point numbers (nearly impossible
to compress), and then with the targeted data sets.
Important suggestions/observations:
-Test reads and writes to understand the asymmetric cost
-Choose representative files - a few VS and PS files
(don't rely just on SMF data sets or SAS data libraries)
-Cost can vary by file - character data usually compresses best
-VSAM cannot used tailored compression dictionaries
-Simple utilities like ICEGENER and REPRO can be used
-Baseline without compression both writes and reads
-Benchmark with compression both writes and reads
-Compare CPU Time difference between baseline and Benchmark
-There are also compaction options in the SYS1.PARMLIB(IGDSMSnn)
and in the DATACLAS definitions that will affect the results,
especially with the new the zEDC compression engine.
6. APAR OA41921 and OA45439 document incompatibility with the Assembly
ENFREQ macro service because OA41921 builds a longer parameter list,
but the issue only exists if MF=L,PLISTVER=MAX was coded, and the
only use of ENFREQ in MXG's ASMTAPEE with MF=L does not have MAX,
so these APARs have no impact on MXG's ASMTAPEE ML-54.
5. APAR PI16853. Obscure. SMF 105 (GDPS Global Mirror) Subtype 2 has
null values in SMF105Lxx LSS section variables if the number of
volumes in the LSS is exactly 256.
4. APAR PI16934 for SMF type 117 records creates a way for you to
choose what part of the actual node name will be written in the
INODNDNM field, which is limited to 32 characters.
3. You can read SMF data directly from the Logger, AND pass filtering
parms using the IFASEXIT with this syntax for your SMF DD:
//SMF DD DSN=IFASMF.DVPLEX.MVS,DISP=SHR,BUFNO=30,
// SUBSYS=(LOGR,IFASEXIT,
// 'FROM=(2014/232,00:00),TO=(2014/232,21:00),LOCAL',
// 'SID(BRVO),SMEP(0100)')
IBM doesn't support splitting the SMF logstreams by LPAR/SYSID, but
you could use this syntax to build PDBs by SYSTEM.
2. APAR OA45174: PROCESSOR CAPACITY CHANGE WITHIN 60 SECONDS OF NEXT
SMF SYNC INTERVAL CAN CAUSE THE NEXT INTERVAL TO BE SKIPPED. The
problem was first observed, when an MXG site started seeing,
occasionally, a 30 minute RMF interval when they should have all
been 15 minute intervals. This APAR corrected the problem, by
changing the granularity of the next interval calculation from one
minute to one second.
1. z/OS 2.1 RED ALERT - APAR PI22817 - DFSORT records out of sequence.
At z/OS 2.1 code levels, DFSORT is intermittently returning records
out of sequence. There is no data loss, but records may be returned
out of sequence to the DFSORT output file. If VERIFY=YES is set at
the installation level, out of sequence conditions are already being
detected. This problem only occurs in z/OS 2.1. No prior releases of
z/OS are impacted.
Please see APAR PI22817 for more details and latest information.
Users Affected:
All z/OS 2.1 DFSORT (HSM1L00) users who SORT data with DFSORT using
the performance path may be affected if there is insufficient
virtual storage below the line at the time of execution. In other
words the potential for error exists for all users of DFSORT SORT
function on z/OS 2.1, but not all users will experience the problem.
Recommended Actions:
Enable VERIFY=YES at the installation level to detect out of
sequence conditions. Affected jobs can be rerun with DEBUG $NOPFP$
to circumvent the issue.
In addition, a ++APAR is available to disable the affected
performance path, as a temporary circumvention, using a new DFSORT
installation option.
Note: Any program that invokes E15 and E35 EXITs to pass records to
DFSORT will not be eligible for the performance path and therefore
not at risk for out of sequence records. DB2 Utilities fall under
this category and are not subject to the issue documented in this
APAR PI22817.
IV. DB2 Technical Notes.
1. Text
V. IMS Technical Notes.
1. Text
VI. SAS Technical Notes.
3. USER ABEND 2008 on z/OS with no other clues; the SAS Log ended after
with the statement %INCLUDE SOURCLIB(BUILDPDB); printed. The cause
was a REGION=0004096K (i.e. 4 MB Region!!) in ancient JCL. Most of
MXG, including enhanced BUILDPDBs, need less than 128 MB, but use of
the ODS Graphics (SGPLOT) can need more region size, but that is
ONLY virtual storage, and NOT physical storage.
2. SAS 9.4 TS1M2 fails with z/OS SYSTEM ABEND 378 followed by USER 999
when the FILEEXIST function is used with a file that DOESN'T exist,
but ONLY IF when you used OPTIONS FILEMSGS; since NOFILEMSGS is the
default, and MXG does NOT set FILEMSGS, and this error does not
occur with NOFILEMSGS.
1. DFSORT APAR PP22817 will NOT impact SAS invoked DFSORTs.
See MVS Technical Note 1., above. IBM notes in the APAR text:
Any program that invokes E15 and E35 EXITs to pass records to DFSORT
will not be eligible for the performance path and therefore not at
risk for out of sequence records. DB2 Utilities fall under this
category and are not subject to the issue documented in this APAR
PI22817.
From SAS Technical Support:
Since all SAS invocations of DFSORT invoke those sort EXits, this
error does NOT affect ANY DFSORT invoked by SAS.
However, you could still have a jobstream that invoked DFSORT in a
stand-alone step that could end up with the SAS code in a subsequent
step not working as expected because the data was not sorted
correctly.
VI.A. WPS Technical Notes.
1. Text
VII. CICS Technical Notes.
1. Text
VIII. Windows NT Technical Notes.
1. Text
IX. z/VM Technical Notes.
1. Text
X. Email notes.
1. Text
XI. Incompatibilities and Installation of MXG vv.yy.
1. Incompatibilities introduced in MXG 32.32 (since MXG 31.31):
See CHANGES.
2. Installation and re-installation procedures are described in detail
in member INSTALL (separate sections for each platform, z/OS, WIN,
or *nix), with examples of common Errors/Warnings messages a new MXG
user might encounter, and in member JCLINSTT for SAS V9.2 or member
JCLINSTW for WPS. INSTALL also shows how to read SMF data on PCs/nix
using the FTP ACCESS METHOD.
XII. Online Documentation of MXG Software.
MXG Documentation is now described in member DOCUMENT.
XIIV. Changes Log
--------------------------Changes Log---------------------------------
You MUST read each Change description to determine if a Change will
impact your site. All changes have been made in this MXG Library.
Member CHANGES always identifies the actual version and release of
MXG Software that is contained in that library.
The CHANGES selection on our homepage at http://www.MXG.com
is always the most current information on MXG Software status,
and is frequently updated.
Important changes are also posted to the MXG-L ListServer, which is
also described by a selection on the homepage. Please subscribe.
The actual code implementation of some changes in MXG SOURCLIB may be
different than described in the change text (which might have printed
only the critical part of the correction that need be made by users).
Scan each source member named in any impacting change for any comments
at the beginning of the member for additional documentation, since the
documentation of new datasets, variables, validation status, and notes,
are often found in comments in the source members.
Alphabetical list of important changes after MXG VV.RR in MXG VV.RR+1:
Dataset/
Member Change Description
See Member CHANGES in your MXG Source Library for THIS MXG version, or
see Member CHANGESS in your MXG Source Library for all 8,935 changes
since 1984, and online at http://www.mxg.com/changes.